Optical amplifiers, specifically Erbium Doped Fiber Amplifiers (EDFA), have introduced a new generation in optical transport systems. EDFAs have allowed the transmission of optical signals over thousands of kilometers without electrical regeneration. The next generation of optical systems is expected to comprise all-optical networks wherein optical signals are routed and switched through the network while remaining in optical form. The path loss of such an optical signal can become large enough to necessitate the use of EDFAs to amplify the signal at various points along the path of travel. Standard EDFAs have been successful in static point to point links where the optical path does not change; however, problems can arise in an all-optical network when the path length and/or path loss changes dynamically.
The natural gain profile of an EDFA is generally nonuniform over the traditional wavelength range of 1530 to 1565 nanometers (nm). In addition, the nonuniformity of the gain profile changes as the input power to the EDFA changes. This effect is commonly called dynamic gain tilt. Since previous optical systems typically utilized static point to point links having constant optical signal power, gain-flattening techniques could be used to compensate for the nonuniform gain profile of an EDFA. When designed for a specific input power and gain profile, these techniques can be used to flatten the gain of an EDFA allowing the full gain window to be utilized in an optical network.
Optical networks that are configured as multichannel or wavelength division multiplexed (WDM) systems benefit from using the full gain window of an optical amplifier. A WDM network divides the available spectrum into channels comprising one or more usable wavelengths, thus allowing multiple signals to be simultaneously transmitted on the same fiber. In WDM systems, it is important to have uniform gain for all channels to limit nonuniform bit error rates (BER) across the channels. In a typical WDM network, gain flatness is in the range of .+-.0.75 dB over the available spectrum. Excess dynamic gain tilt can cause significant BERs on low gain channels in a cascade of optical amplifiers.
In order to increase the channel count in a WDM system and still maintain a flat gain profile, techniques for flattening the gain of an EDFA have been developed. One technique is to filter the optical signal as disclosed in, Erbium-doped fiber amplifier flattened beyond 40 nm using long-period grating, authored by P. F. Wysocki, J. Judkins, R. Espindola, M. Andrejco, A Vengsarkar and K. Walker and published in Optical Fiber Communication Conference, Optical Society of America, Washington, D.C., paper PD2, 1997. Gain flattening may also be accomplished by changing the doping profile within the fiber so that the natural gain shape is flatter as disclosed in, 1.5 um broadband amplification by tellurite-based EDFAs, authored by A. Mori, Y. Ohishi, M. Yamada, H. Ono, Y. Nishida, K. Oikawa, and S. Sudo and published in Optical Fiber Communication Conference, Optical Society of America, Washington, D.C., paper PD1, 1997. Unfortunately, these techniques are generally only applicable to cases where the input power to the EDFA is constant or can change by only a small amount. When the input power to the EDFA changes by a large amount, dynamic gain tilt can still become a problem.
Recently, techniques have been investigated where the dynamic range of the EDFA is expanded with the operating gain and gain flatness being maintained. One of these techniques involves using a loop controller to lock the gain of the EDFA as disclosed in Optical Amplifiers and their Applications, authored by A. K. Sravistava, et al., and published in Vol. V of OSA Trends in Optics and Photonics Series, Optical Society of America, Washington, D.C., paper PDP4, 1996. In this approach, the input and output optical power of the EDFA are sampled and control signals to the EDFA are adjusted to keep the ratio of the input power to output power constant. Another technique involves using an acousto-optic tunable filter at the midstage point of an EDFA as disclosed in, Dynamic gain equalization of erbium-doped fiber amplifier with all-fiber acousto-optic tunable filters, authored by H. S. Kim, S. H. Yun, H. K. Kim, N. Park, B. and Y. Kim, and published in Optical Fiber Communication Conference, Optical Society of America, Washington, D.C., paper WG4, 1998. In this approach, the filter shape is tailored to maximize the gain flatness for a given input power and gain shape. Still another technique involves using a variable optical attenuator (VOA) between two successive gain stages as disclosed in, WDM linear repeater gain control scheme by automatic maximum power channel selection for photonic transport network, authored by N. Takachio, H. Suzuki, M. Koga, and O. Ishida, and published in Optical Fiber Communication Conference, Optical Society of America, Washington, D.C., paper WJ4, 1998. The disclosed technique improves the dynamic range capability slightly but has only been demonstrated over a narrow wavelength range of 1540 to 1560 nm and does not use any external network information in determining the ideal operating conditions for the module.
The above techniques help to solve the problem of maintaining the gain flatness by gain locking the amplifier, however utilizing the above techniques, changes in the total output power of the EDFA will still occur as the input power changes. The ideal case for a dynamic network is to have a constant output power per channel at the output of each optical amplifier for a wide range of input powers and varying numbers of channels. Gain locking techniques on their own maintain the gain flatness of the amplifier but can only do this over a limited input range and by design, do not maintain constant output power. A constant output power can limit nonlinear effects in the fiber in excessive output power cases and prevents downstream components from experiencing large changes in received optical power. As a result, changes in power levels can be localized in the network.
This presents a problem in all-optical networks since the optical power input to an EDFA will vary dramatically due primarily to two effects. The first effect is the change brought about by a switching event required for restoring traffic in the network. Optical signals originate at different sources and travel different paths to get to the EDFA. In an all-optical network, the light launched into one fiber may be re-routed to another fiber if the first fiber's connection is broken. In the worst case, the two paths will have extremely different loss characteristics that will result in a large variation of input power to the receiving EDFA. The second effect results from changes in network configuration, such as, changing the total number of channels in a WDM network's configuration. When the individual power per channel is constant, changing the number of channels in a WDM system changes the total power in the fiber. EDFAs experience saturation effects based on the total optical power in the fiber. Thus, changes in fiber loss and network configuration can lead to dynamic changes to the input power at an EDFA resulting in signal degradation due to gain tilt.
Therefore, what is needed is a way to amplify an optical signal in an optical network wherein the optical output power per channel remains constant regardless of the dynamic variations of the input signal power which may result from fiber loss and/or network reconfigurations. This will maintain channel integrity over the full range of network re-configurations.